A shovel is pushed into dry soil and lifted again. Each insertion lands with a dull thud followed by gritty earth shuffling.
Seedance 2.0ByteDance · proprietary
Kling 3.0 OmniKuaishou · proprietary
Veo 3.1Google · proprietary
LTX-2.3Lightricks · open
Ovi 1.1character.ai · open
JavisDiT++JavisDiT · open
MagiHumanGAIR-NLP · open
Per-prompt rubric
Basic standards
Visual presence
- a shovel
- dry soil
Event: a shovel is pushed into dry soil and lifted
Audio presence
- a shovel
- dry soil
Sound: shovel cutting into dry soil
Key standards
Visual physical commonsense (V-PC)
- A shovel is pushed into dry soil and lifted again, with visible soil displacement on the shovel blade.
Audio physical commonsense (A-PC)
- Each insertion has a dull thud followed by gritty earth-shuffling timbre consistent with a metal shovel cutting into dry granular soil.
Cross-modal physical commonsense (AV-PC)
- Each thud onset aligns in time with a visible insertion of the shovel into the soil.